diff mbox

Set NODELETE flag after checking for NULL pointer

Message ID 1457721020-5824-1-git-send-email-aurelien@aurel32.net
State New, archived
Headers

Commit Message

Aurelien Jarno March 11, 2016, 6:30 p.m. UTC 
  The commit b632bdd3 moved the setting of the DF_1_NODELETE flag earlier
in the dl_open_worker function. However when calling dlopen with both
RTLD_NODELETE and RTLD_NOLOAD (which in practice also requires
RTLD_LAZY), the pointer returned by _dl_map_object is NULL. This
condition is checked just after setting the flag, while it should be
done before. Fix that.

Changelog:
	[BZ #19810]
	* elf/dl-open.c (dl_open_worker): Set DF_1_NODELETE flag later.
---
 ChangeLog     |  5 +++++
 elf/dl-open.c | 12 ++++++------
 2 files changed, 11 insertions(+), 6 deletions(-)

Comments

Florian Weimer March 11, 2016, 6:33 p.m. UTC | #1 
On 03/11/2016 07:30 PM, Aurelien Jarno wrote:
> The commit b632bdd3 moved the setting of the DF_1_NODELETE flag earlier
> in the dl_open_worker function. However when calling dlopen with both
> RTLD_NODELETE and RTLD_NOLOAD (which in practice also requires
> RTLD_LAZY), the pointer returned by _dl_map_object is NULL. This
> condition is checked just after setting the flag, while it should be
> done before. Fix that.

How hard is it to create a test case for this?

Thanks,
Florian
Aurelien Jarno March 11, 2016, 6:43 p.m. UTC | #2 
On 2016-03-11 19:33, Florian Weimer wrote:
> On 03/11/2016 07:30 PM, Aurelien Jarno wrote:
> > The commit b632bdd3 moved the setting of the DF_1_NODELETE flag earlier
> > in the dl_open_worker function. However when calling dlopen with both
> > RTLD_NODELETE and RTLD_NOLOAD (which in practice also requires
> > RTLD_LAZY), the pointer returned by _dl_map_object is NULL. This
> > condition is checked just after setting the flag, while it should be
> > done before. Fix that.
> 
> How hard is it to create a test case for this?

That should be relatively easy I guess. If the patch makes sense, I can
try to add one.

Aurelien
Florian Weimer March 11, 2016, 6:56 p.m. UTC | #3 
On 03/11/2016 07:43 PM, Aurelien Jarno wrote:
> On 2016-03-11 19:33, Florian Weimer wrote:
>> On 03/11/2016 07:30 PM, Aurelien Jarno wrote:
>>> The commit b632bdd3 moved the setting of the DF_1_NODELETE flag earlier
>>> in the dl_open_worker function. However when calling dlopen with both
>>> RTLD_NODELETE and RTLD_NOLOAD (which in practice also requires
>>> RTLD_LAZY), the pointer returned by _dl_map_object is NULL. This
>>> condition is checked just after setting the flag, while it should be
>>> done before. Fix that.
>>
>> How hard is it to create a test case for this?
> 
> That should be relatively easy I guess. If the patch makes sense, I can
> try to add one.

The change looks reasonable and correct to me, but the dynamic linker
scares me …

Thanks,
Florian
diff mbox

Patch

diff --git a/ChangeLog b/ChangeLog
index 440b021..fe113b7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,8 @@ 
+2016-03-09  Aurelien Jarno  <aurelien@aurel32.net>
+
+	[BZ #19810]
+	* elf/dl-open.c (dl_open_worker): Set DF_1_NODELETE flag later.
+
 2016-03-11  Rajalakshmi Srinivasaraghavan  <raji@linux.vnet.ibm.com>
 
 	* sysdeps/powerpc/powerpc32/power4/memcmp.S (memcmp): Rearrange
diff --git a/elf/dl-open.c b/elf/dl-open.c
index 6f178b3..3e5df48 100644
--- a/elf/dl-open.c
+++ b/elf/dl-open.c
@@ -226,12 +226,6 @@  dl_open_worker (void *a)
   args->map = new = _dl_map_object (call_map, file, lt_loaded, 0,
 				    mode | __RTLD_CALLMAP, args->nsid);
 
-  /* Mark the object as not deletable if the RTLD_NODELETE flags was passed.
-     Do this early so that we don't skip marking the object if it was
-     already loaded.  */
-  if (__glibc_unlikely (mode & RTLD_NODELETE))
-    new->l_flags_1 |= DF_1_NODELETE;
-
   /* If the pointer returned is NULL this means the RTLD_NOLOAD flag is
      set and the object is not already loaded.  */
   if (new == NULL)
@@ -240,6 +234,12 @@  dl_open_worker (void *a)
       return;
     }
 
+  /* Mark the object as not deletable if the RTLD_NODELETE flags was passed.
+     Do this early so that we don't skip marking the object if it was
+     already loaded.  */
+  if (__glibc_unlikely (mode & RTLD_NODELETE))
+    new->l_flags_1 |= DF_1_NODELETE;
+
   if (__glibc_unlikely (mode & __RTLD_SPROF))
     /* This happens only if we load a DSO for 'sprof'.  */
     return;